BEGIN:VCALENDAR
VERSION:2.0
PRODID:icalendar-ruby
CALSCALE:GREGORIAN
X-WR-CALNAME:CBE Seminar: Yuhe Tian\, West Virginia University
X-WR-TIMEZONE:Eastern Time (US & Canada)
BEGIN:VEVENT
DTSTAMP:20260513T221455Z
UID:tag:localist.com\,2008:EventInstance_52002010320926
DTSTART:20260306T140000Z
DTEND:20260306T150000Z
DESCRIPTION:Advanced Process Control and Real-Time Decision-Making via Rein
 forcement Learning\n\n \n\nReinforcement Learning (RL) has significantly e
 xpanded the field of process control and real-time decision-making in the 
 past decade\, enabling the optimal operation of highly complex dynamic sys
 tems via the use of neural networks. RL has been exceptionally useful in s
 ituations where it is challenging to create a mechanistic model or develop
  a high-accuracy data-driven surrogate representation. RL has proven to be
  transformative in the areas of robotics\, game-playing\, and hierarchical
  decision-making\, but has been challenging to implement for the direct co
 ntrol of chemical process systems. In this talk\, we will first provide an
  overview on the basics of neural networks and state-of-the-art RL algorit
 hms. Then we will focus on two key challenges in RL-based chemical process
  control: (i) how to improve the learning efficiency which typically requi
 res a very large amount of training data until making reliable decisions? 
 and (ii) how to ensure the safety and stability of the learned control pol
 icy? To address these challenges\, we will introduce a novel transfer-lear
 ned RL algorithm that leverages Y-wise Affine Neural Networks. This specia
 lized neural network architecture can exactly represent the explicit contr
 ol policy from multi-parametric model predictive control (mp-MPC). The Y-w
 ise Affine Neural Networks can thus serve as a hot start and transfer the 
 mp-MPC knowledge to RL training. This contributes to fully eliminate the u
 nsafe and time-consuming exploration stage during RL training\, while prov
 iding control actions with confidence. We will also discuss the implementa
 tion of continuous policy improvement to heuristically guarantee that the 
 mp-MPC solution serves as an effective lower bound to the transfer-learned
  RL. The computational and practical advantages of this algorithm will be 
 demonstrated on the control and real-time decision-making of multiple repr
 esentative chemical process systems.
GEO:39.680031;-75.749885
LOCATION:Colburn Lab\, 102
SUMMARY:CBE Seminar: Yuhe Tian\, West Virginia University
URL;VALUE=URI:https://events.udel.edu/event/cbe-seminar-yuhe-tian-west-virg
 inia-university
CATEGORIES:Academics
CATEGORIES:College of Engineering
END:VEVENT
END:VCALENDAR